Publications (68)
Meta-Learning for Low-resource Natural Language Generation in Task-oriented Dialogue Systems
Fei Mi, Minlie Huang, Jiyong Zhang +1
Natural language generation (NLG) is an essential component of task-oriented dialogue systems. Despite the recent success of neural approaches for NLG, they are typically developed…
UAlign: Leveraging Uncertainty Estimations for Factuality Alignment on Large Language Models
Boyang Xue, Fei Mi, Qi Zhu +6
Despite demonstrating impressive capabilities, Large Language Models (LLMs) still often struggle to accurately express the factual knowledge they possess, especially in cases where…
TPE: Towards Better Compositional Reasoning over Conceptual Tools with Multi-persona Collaboration
Hongru Wang, Huimin Wang, Lingzhi Wang +6
Large language models (LLMs) have demonstrated exceptional performance in planning the use of various functional tools, such as calculators and retrievers, particularly in question…
Enhancing Large Language Models Against Inductive Instructions with Dual-critique Prompting
Rui Wang, Hongru Wang, Fei Mi +4
Numerous works are proposed to align large language models (LLMs) with human intents to better fulfill instructions, ensuring they are trustful and helpful. Nevertheless, some huma…
EnvFactory: Scaling Tool-Use Agents via Executable Environments Synthesis and Robust RL
Minrui Xu, Zilin Wang, Mengyi DENG +12
Equipping LLMs with tool-use capabilities via Agentic Reinforcement Learning (Agentic RL) is bottlenecked by two challenges: the lack of scalable, robust execution environments and…
ADER: Adaptively Distilled Exemplar Replay Towards Continual Learning for Session-based Recommendation
Fei Mi, Xiaoyu Lin, Boi Faltings
Session-based recommendation has received growing attention recently due to the increasing privacy concern. Despite the recent success of neural session-based recommenders, they ar…
DAST: Difficulty-Aware Self-Training on Large Language Models
Boyang Xue, Qi Zhu, Hongru Wang +8
Present Large Language Models (LLM) self-training methods always under-sample on challenging queries, leading to inadequate learning on difficult problems which limits LLMs' abilit…
CoSafe: Evaluating Large Language Model Safety in Multi-Turn Dialogue Coreference
Erxin Yu, Jing Li, Ming Liao +4
As large language models (LLMs) constantly evolve, ensuring their safety remains a critical research problem. Previous red-teaming approaches for LLM safety have primarily focused…
CINS: Comprehensive Instruction for Few-shot Learning in Task-oriented Dialog Systems
Fei Mi, Yitong Li, Yasheng Wang +2
As labeling cost for different modules in task-oriented dialog (ToD) systems is high, a major challenge in practice is to learn different tasks with the least amount of labeled dat…
Gaining Wisdom from Setbacks: Aligning Large Language Models via Mistake Analysis
Kai Chen, Chunwei Wang, Kuo Yang +11
The rapid development of large language models (LLMs) has not only provided numerous opportunities but also presented significant challenges. This becomes particularly evident when…
InfMem: Learning System-2 Memory Control for Long-Context Agent
Xinyu Wang, Mingze Li, Peng Lu +6
Reasoning over ultra-long documents requires synthesizing sparse evidence scattered across distant segments under strict memory constraints. While streaming agents enable scalable…
The Synergy Dilemma of Long-CoT SFT and RL: Investigating Post-Training Techniques for Reasoning VLMs
Jierun Chen, Tiezheng Yu, Haoli Bai +11
Large vision-language models (VLMs) increasingly adopt post-training techniques such as long chain-of-thought (CoT) supervised fine-tuning (SFT) and reinforcement learning (RL) to…
UniMS-RAG: A Unified Multi-source Retrieval-Augmented Generation for Personalized Dialogue Systems
Hongru Wang, Wenyu Huang, Yang Deng +6
Large Language Models (LLMs) has shown exceptional capabilities in many natual language understanding and generation tasks. However, the personalization issue still remains a much-…
You Live More Than Once: Towards Hierarchical Skill Meta-Evolving
Xujun Li, Kehan Zheng, Mingyuan Zhao +7
Test-time skill evolving is regarded as a new paradigm for enhancing deployed agentic systems. Existing works mainly focus on hard-coded skill evolving strategies or parametric lea…
Modeling Complex Dialogue Mappings via Sentence Semantic Segmentation Guided Conditional Variational Auto-Encoder
Bin Sun, Shaoxiong Feng, Yiwei Li +4
Complex dialogue mappings (CDM), including one-to-many and many-to-one mappings, tend to make dialogue models generate incoherent or dull responses, and modeling these mappings rem…
Personalization in Goal-Oriented Dialog
Chaitanya K. Joshi, Fei Mi, Boi Faltings
The main goal of modeling human conversation is to create agents which can interact with people in both open-ended and goal-oriented scenarios. End-to-end trained neural dialog sys…
ReliableMath: Benchmark of Reliable Mathematical Reasoning on Large Language Models
Boyang Xue, Qi Zhu, Rui Wang +8
Although demonstrating remarkable performance on reasoning tasks, Large Language Models (LLMs) still tend to fabricate unreliable responses when confronted with problems that are u…
Representation Memorization for Fast Learning New Knowledge without Forgetting
Fei Mi, Tao Lin, Boi Faltings
The ability to quickly learn new knowledge (e.g. new classes or data distributions) is a big step towards human-level intelligence. In this paper, we consider scenarios that requir…
EssayBench: Evaluating Large Language Models in Multi-Genre Chinese Essay Writing
Fan Gao, Dongyuan Li, Ding Xia +4
Chinese essay writing and its evaluation are critical in educational contexts, yet the capabilities of Large Language Models (LLMs) in this domain remain largely underexplored. Exi…
Self-training Improves Pre-training for Few-shot Learning in Task-oriented Dialog Systems
Fei Mi, Wanhao Zhou, Fengyu Cai +3
As the labeling cost for different modules in task-oriented dialog (ToD) systems is expensive, a major challenge is to train different modules with the least amount of labeled data…
Group Pattern Selection Optimization: Let LRMs Pick the Right Pattern for Reasoning
Hanbin Wang, Jingwei Song, Jinpeng Li +2
Large reasoning models (LRMs) exhibit diverse high-level reasoning patterns (e.g., direct solution, reflection-and-verification, and exploring multiple solutions), yet prevailing t…
Improving Factual Consistency for Knowledge-Grounded Dialogue Systems via Knowledge Enhancement and Alignment
Boyang Xue, Weichao Wang, Hongru Wang +7
Pretrained language models (PLMs) based knowledge-grounded dialogue systems are prone to generate responses that are factually inconsistent with the provided knowledge source. In s…
UniRetriever: Multi-task Candidates Selection for Various Context-Adaptive Conversational Retrieval
Hongru Wang, Boyang Xue, Baohang Zhou +5
Conversational retrieval refers to an information retrieval system that operates in an iterative and interactive manner, requiring the retrieval of various external resources, such…
Teaching Large Reasoning Models Effective Reflection
Hanbin Wang, Jingwei Song, Jinpeng Li +5
Large Reasoning Models (LRMs) have recently shown impressive performance on complex reasoning tasks, often by engaging in self-reflective behaviors such as self-critique and backtr…
Pangu Ultra: Pushing the Limits of Dense Large Language Models on Ascend NPUs
Yichun Yin, Wenyong Huang, Kaikai Song +49
We present Pangu Ultra, a Large Language Model (LLM) with 135 billion parameters and dense Transformer modules trained on Ascend Neural Processing Units (NPUs). Although the field…
Exploring Effective Information Utilization in Multi-Turn Topic-Driven Conversations
Jiatong Li, Bin He, Fei Mi
Conversations are always related to certain topics. However, it is challenging to fuse dialogue history and topic information from various sources at the same time in current dialo…
FollowBench: A Multi-level Fine-grained Constraints Following Benchmark for Large Language Models
Yuxin Jiang, Yufei Wang, Xingshan Zeng +7
The ability to follow instructions is crucial for Large Language Models (LLMs) to handle various real-world applications. Existing benchmarks primarily focus on evaluating pure res…
Pan More Gold from the Sand: Refining Open-domain Dialogue Training with Noisy Self-Retrieval Generation
Yihe Wang, Yitong Li, Yasheng Wang +6
Real human conversation data are complicated, heterogeneous, and noisy, from which building open-domain dialogue systems remains a challenging task. In fact, such dialogue data sti…
ClusterUCB: Efficient Gradient-Based Data Selection for Targeted Fine-Tuning of LLMs
Zige Wang, Qi Zhu, Fei Mi +3
Gradient-based data influence approximation has been leveraged to select useful data samples in the supervised fine-tuning of large language models. However, the computation of gra…
Continual Learning for Natural Language Generation in Task-oriented Dialog Systems
Fei Mi, Liangwei Chen, Mengjie Zhao +2
Natural language generation (NLG) is an essential component of task-oriented dialog systems. Despite the recent success of neural approaches for NLG, they are typically developed i…
DecompEval: Evaluating Generated Texts as Unsupervised Decomposed Question Answering
Pei Ke, Fei Huang, Fei Mi +4
Existing evaluation metrics for natural language generation (NLG) tasks face the challenges on generalization ability and interpretability. Specifically, most of the well-performed…
KDRL: Post-Training Reasoning LLMs via Unified Knowledge Distillation and Reinforcement Learning
Hongling Xu, Qi Zhu, Heyuan Deng +6
Recent advances in large language model (LLM) post-training have leveraged two distinct paradigms to enhance reasoning capabilities: reinforcement learning (RL) and knowledge disti…
Self-Error-Instruct: Generalizing from Errors for LLMs Mathematical Reasoning
Erxin Yu, Jing Li, Ming Liao +7
Although large language models demonstrate strong performance across various domains, they still struggle with numerous bad cases in mathematical reasoning. Previous approaches to…
KPT: Keyword-guided Pre-training for Grounded Dialog Generation
Qi Zhu, Fei Mi, Zheng Zhang +6
Incorporating external knowledge into the response generation process is essential to building more helpful and reliable dialog agents. However, collecting knowledge-grounded conve…
SELF: Self-Evolution with Language Feedback
Jianqiao Lu, Wanjun Zhong, Wenyong Huang +9
Large Language Models (LLMs) have demonstrated remarkable versatility across various domains. To further advance LLMs, we propose 'SELF' (Self-Evolution with Language Feedback), a…
Role Prompting Guided Domain Adaptation with General Capability Preserve for Large Language Models
Rui Wang, Fei Mi, Yi Chen +5
The growing interest in Large Language Models (LLMs) for specialized applications has revealed a significant challenge: when tailored to specific domains, LLMs tend to experience c…
Aligning Large Language Models with Human: A Survey
Yufei Wang, Wanjun Zhong, Liangyou Li +6
Large Language Models (LLMs) trained on extensive textual corpora have emerged as leading solutions for a broad array of Natural Language Processing (NLP) tasks. Despite their nota…
YODA: Teacher-Student Progressive Learning for Language Models
Jianqiao Lu, Wanjun Zhong, Yufei Wang +10
Although large language models (LLMs) have demonstrated adeptness in a range of tasks, they still lag behind human learning efficiency. This disparity is often linked to the inhere…
Towards Diverse, Relevant and Coherent Open-Domain Dialogue Generation via Hybrid Latent Variables
Bin Sun, Yitong Li, Fei Mi +3
Conditional variational models, using either continuous or discrete latent variables, are powerful for open-domain dialogue response generation. However, previous works show that c…
UniDS: A Unified Dialogue System for Chit-Chat and Task-oriented Dialogues
Xinyan Zhao, Bin He, Yasheng Wang +6
With the advances in deep learning, tremendous progress has been made with chit-chat dialogue systems and task-oriented dialogue systems. However, these two systems are often tackl…
Pangu Embedded: An Efficient Dual-system LLM Reasoner with Metacognition
Hanting Chen, Yasheng Wang, Kai Han +21
This work presents Pangu Embedded, an efficient Large Language Model (LLM) reasoner developed on Ascend Neural Processing Units (NPUs), featuring flexible fast and slow thinking ca…
Cue-CoT: Chain-of-thought Prompting for Responding to In-depth Dialogue Questions with LLMs
Hongru Wang, Rui Wang, Fei Mi +5
Large Language Models (LLMs), such as \texttt{ChatGPT}, greatly empower dialogue systems with strong language understanding and generation capabilities. However, most of the previo…
Defending Large Language Models Against Jailbreaking Attacks Through Goal Prioritization
Zhexin Zhang, Junxiao Yang, Pei Ke +3
While significant attention has been dedicated to exploiting weaknesses in LLMs through jailbreaking attacks, there remains a paucity of effort in defending against these attacks.…
ReSee: Responding through Seeing Fine-grained Visual Knowledge in Open-domain Dialogue
Haoqin Tu, Yitong Li, Fei Mi +1
Incorporating visual knowledge into text-only dialogue systems has become a potential direction to imitate the way humans think, imagine, and communicate. However, existing multimo…
Pangu Pro MoE: Mixture of Grouped Experts for Efficient Sparsity
Yehui Tang, Xiaosong Li, Fangcheng Liu +19
The surgence of Mixture of Experts (MoE) in Large Language Models promises a small price of execution cost for a much larger model parameter count and learning capacity, because on…
SynCoBERT: Syntax-Guided Multi-Modal Contrastive Pre-Training for Code Representation
Xin Wang, Yasheng Wang, Fei Mi +7
Code representation learning, which aims to encode the semantics of source code into distributed vectors, plays an important role in recent deep-learning-based models for code inte…
Mixture of insighTful Experts (MoTE): The Synergy of Thought Chains and Expert Mixtures in Self-Alignment
Zhili Liu, Yunhao Gou, Kai Chen +8
As the capabilities of large language models (LLMs) continue to expand, aligning these models with human values remains a significant challenge. Recent studies show that reasoning…
RUBAS: Rubric-Based Reinforcement Learning for Agent Safety
Xian Qi Loye, Qinglin Su, Zhexin Zhang +5
The evolution of LLMs into tool-enabled agents creates a new class of safety challenges associated with real-world execution rather than simple text generation. Existing alignment…
How Should We Enhance the Safety of Large Reasoning Models: An Empirical Study
Zhexin Zhang, Xian Qi Loye, Victor Shea-Jay Huang +8
Large Reasoning Models (LRMs) have achieved remarkable success on reasoning-intensive tasks such as mathematics and programming. However, their enhanced reasoning capabilities do n…
Dynamic Stochastic Decoding Strategy for Open-Domain Dialogue Generation
Yiwei Li, Fei Mi, Yitong Li +4
Stochastic sampling strategies such as top-k and top-p have been widely used in dialogue generation task. However, as an open-domain chatting system, there will be two different co…
Masking as an Efficient Alternative to Finetuning for Pretrained Language Models
Mengjie Zhao, Tao Lin, Fei Mi +2
We present an efficient method of utilizing pretrained language models, where we learn selective binary masks for pretrained weights in lieu of modifying them through finetuning. E…
Towards Identifying Social Bias in Dialog Systems: Frame, Datasets, and Benchmarks
Jingyan Zhou, Jiawen Deng, Fei Mi +6
The research of open-domain dialog systems has been greatly prospered by neural models trained on large-scale corpora, however, such corpora often introduce various safety problems…
MoralDial: A Framework to Train and Evaluate Moral Dialogue Systems via Moral Discussions
Hao Sun, Zhexin Zhang, Fei Mi +6
Morality in dialogue systems has raised great attention in research recently. A moral dialogue system aligned with users' values could enhance conversation engagement and user conn…
SLIM: Explicit Slot-Intent Mapping with BERT for Joint Multi-Intent Detection and Slot Filling
Fengyu Cai, Wanhao Zhou, Fei Mi +1
Utterance-level intent detection and token-level slot filling are two key tasks for natural language understanding (NLU) in task-oriented systems. Most existing approaches assume t…
Large Language Models as Source Planner for Personalized Knowledge-grounded Dialogue
Hongru Wang, Minda Hu, Yang Deng +7
Open-domain dialogue system usually requires different sources of knowledge to generate more informative and evidential responses. However, existing knowledge-grounded dialogue sys…
Memory-T1: Reinforcement Learning for Temporal Reasoning in Multi-session Agents
Yiming Du, Baojun Wang, Yifan Xiang +11
Temporal reasoning over long, multi-session dialogues is a critical capability for conversational agents. However, existing works and our pilot study have shown that as dialogue hi…
PanGu-Bot: Efficient Generative Dialogue Pre-training from Pre-trained Language Model
Fei Mi, Yitong Li, Yulong Zeng +7
In this paper, we introduce PanGu-Bot, a Chinese pre-trained open-domain dialogue generation model based on a large pre-trained language model (PLM) PANGU-alpha (Zeng et al.,2021).…
Constructing Highly Inductive Contexts for Dialogue Safety through Controllable Reverse Generation
Zhexin Zhang, Jiale Cheng, Hao Sun +5
Large pretrained language models can easily produce toxic or biased content, which is prohibitive for practical use. In order to detect such toxic generations, existing methods rel…
Context Tree for Adaptive Session-based Recommendation
Fei Mi, Boi Faltings
There has been growing interests in recent years from both practical and research perspectives for session-based recommendation tasks as long-term user profiles do not often exist…
COLD: A Benchmark for Chinese Offensive Language Detection
Jiawen Deng, Jingyan Zhou, Hao Sun +4
Offensive language detection is increasingly crucial for maintaining a civilized social media platform and deploying pre-trained language models. However, this task in Chinese is s…
Rethinking Expert Trajectory Utilization in LLM Post-training for Mathematical Reasoning
Bowen Ding, Yuhan Chen, Jiayang Lyv +9
Supervised Fine-Tuning (SFT) and Reinforcement Learning (RL) dominate the post-training landscape for mathematical reasoning, yet differ fundamentally in their reliance on expert t…
Compilable Neural Code Generation with Compiler Feedback
Xin Wang, Yasheng Wang, Yao Wan +7
Automatically generating compilable programs with (or without) natural language descriptions has always been a touchstone problem for computational linguistics and automated softwa…
Memory Augmented Neural Model for Incremental Session-based Recommendation
Fei Mi, Boi Faltings
Increasing concerns with privacy have stimulated interests in Session-based Recommendation (SR) using no personal data other than what is observed in the current browser session. E…
ToolACE-MT: Non-Autoregressive Generation for Agentic Multi-Turn Interaction
Xingshan Zeng, Weiwen Liu, Lingzhi Wang +6
Agentic task-solving with Large Language Models (LLMs) requires multi-turn, multi-step interactions, often involving complex function calls and dynamic user-agent exchanges. Existi…
Continual Prompt Tuning for Dialog State Tracking
Qi Zhu, Bing Li, Fei Mi +2
A desirable dialog system should be able to continually learn new skills without forgetting old ones, and thereby adapt to new domains or tasks in its life cycle. However, continua…
LMTurk: Few-Shot Learners as Crowdsourcing Workers in a Language-Model-as-a-Service Framework
Mengjie Zhao, Fei Mi, Yasheng Wang +4
Vast efforts have been devoted to creating high-performance few-shot learners, i.e., large-scale pretrained language models (PLMs) that perform well with little downstream task tra…
Entropy Centroids as Intrinsic Rewards for Test-Time Scaling
Wenshuo Zhao, Qi Zhu, Xingshan Zeng +4
An effective way to scale up test-time compute of large language models is to sample multiple responses and then select the best one, as in Grok Heavy and Gemini Deep Think. Existi…
Data Management For Training Large Language Models: A Survey
Zige Wang, Wanjun Zhong, Yufei Wang +6
Data plays a fundamental role in training Large Language Models (LLMs). Efficient data management, particularly in formulating a well-suited training dataset, is significant for en…